Two frame modes
Animate forward from one still, or set a first and a last frame and let the clip travel between them — the mode to use when a shot has to land on a logo or a held pose.
Image to video
Free AI image to video generator
Upload a photo, describe the motion, and get a real clip back — 5 to 15 seconds, landscape or portrait, 480p through 4K. A free account covers your first one, no card.
One render
What you get back
Source
A photo you upload, or a still you just generated. JPEG, PNG or WebP up to 10 MB.
Motion
One or two sentences of direction — the camera move, and the one thing that should happen.
Format
Landscape 16:9 or portrait 9:16, decided before the render rather than cropped after.
Length
5 to 15 seconds per clip, set on a slider.
Quality
480p while you are drafting, up to 4K for the take you keep.
How it works
Three steps, one still
Image-to-video buys the look outright. The picture settles composition, colour and identity before a single second is rendered, so the only thing left to get right is the motion.
- Step 1
Bring the still
Drag a photo in, or animate an image you generated in the Image Studio moments earlier. The studio recognises its own work, so a generated still goes straight to motion with no identity check. Frame it the way the clip should open: leave room for the move, and keep the subject off the extreme edges, because edges warp first.
- Step 2
Describe only what moves
The engine can already see the picture, so re-describing it wastes the prompt and sometimes fights it. Name the camera move and the single thing that should happen. A motion style picker sets the energy if you would rather not write it out.
- Step 3
Pick the shape and render
Choose landscape or portrait, a length between 5 and 15 seconds, and a quality rung. Sound is generated with the picture and can be switched off. The finished clip lands in your Library.
What you control
Enough control to hit the shot
The still locks the look. Everything here decides what the camera does with it, and what comes out the other end.
Landscape or portrait
16:9 for YouTube and the web, 9:16 for Shorts, Reels and TikTok. You pick the shape before the render, so a vertical clip is generated vertical rather than cropped down to it.
5 to 15 seconds
Short takes first. Motion quality decays as a generation runs long, and a 5-second re-roll is the cheapest way to fix a shot that missed.
480p through 4K
Draft at 480p while you are still hunting for the motion, then re-run the take that worked at a higher rung. Every control behaves identically at all four.
Audio, or silence
Sound is generated together with the picture. Leave it on for atmosphere, or switch it off when the clip is heading under a music bed on the timeline.
Frames stay pure
In the frame modes your still is the only image the render sees — no extra reference images ride along with it. That is precisely what keeps the subject looking like itself.
The honest version
What free actually buys
Every second of generated video costs real GPU time, so nobody renders it without limit for nothing. Here is exactly what a free account covers, and what it does not.
- 50 starter credits, no card. That is one 5-second clip at 480p (40 credits), or 4 images to animate later.
- 480p bills at 8 credits a second, so a 5-second draft is 40. The price is on screen before you press generate.
- No plan watermark. The only overlay PrismPoster burns into a clip is the AI-disclosure label required on real-person likeness work, and that is identical on every plan.
When a tool advertises unlimited free AI video, it is recovering the cost somewhere: a queue you wait in, a watermark, a hard resolution cap, or rights over what you make. We would rather publish the rate.
After the clip
One clip is not a video
A finished piece is several animated stills cut together. That happens in the same place, on the same credits, without moving your footage to another tool.
A timeline that generates
Clips, images, audio and titles on one multi-track timeline, with the generators inside the editor so a new take lands at the playhead. Exports render server-side as MP4 at any length.
Everything lands in the Library
Each still and clip stays reusable: animate the same image again with different motion, or pull an earlier take back onto the timeline when the new one is worse.
Labelled the way the law expects
Every image, video and audio export carries embedded machine-readable provenance metadata (IPTC, XMP, EXIF, MP4, ID3), aligned to EU AI Act Article 50(2). Video rendered on the timeline also carries a C2PA signature. No export carries a visible badge, so telling your audience at publication stays yours.
Questions
Before you upload anything
Is the AI image to video generator free?
Yes, to a point. A free account gets 50 starter credits with no card — enough for one 5-second clip at 480p (40 credits), or 4 images you can animate later. Paid plans add monthly credits, which is what puts the 1080p and 4K rungs within reach.
Is there an unlimited free image to video AI?
Not honestly. Every second of generated video costs GPU time, so a tool promising unlimited free rendering is recovering that cost elsewhere — a queue, a watermark, a hard resolution cap, or rights over what you create. PrismPoster publishes the rate instead: 480p costs 8 credits a second, and the price is on screen before you generate.
What images can I upload?
JPEG, PNG and WebP up to 10 MB, by drag and drop or file picker. You can also skip the upload entirely and animate an image you just generated in the Image Studio, which costs 12 credits and lets reference images keep a whole set of stills on-model before any of them move.
How long can the video be?
Each render is 5 to 15 seconds. Longer pieces are assembled rather than generated: cut several clips together on the timeline editor and export one MP4, which has no length limit.
Can I animate a photo of a real person?
Not by uploading their photo — the video engine rejects raw photographs of identifiable people, deliberately. A photoreal character you generated in the Image Studio is different: the studio knows that image is its own work, so it animates with no identity check. Putting an actual person on screen runs through Creator Cast instead: that person sets up a consented likeness from their own photos, can pause or revoke it at any time, and the likeness is what renders.
Does PrismPoster post the video to social media?
No. PrismPoster creates and finishes the media; you export it and post it yourself. There is no scheduler and no connected posting — the studio ends at a file you own.
Explore more
Image-to-video guide
The full workflow: making a still worth animating, prompting motion instead of scene, and the failure modes to dodge.
Text-to-video vs image-to-video
When to describe a scene from scratch, and when to hand the engine a finished picture instead.
Which resolution to pay for
What 480p, 720p, 1080p and 4K are each actually good for, and when the jump is worth it.
Everything in the studio
Image, video, music, Storyboard and the timeline editor, on one set of credits.
Credits and plans
What a clip costs at each quality rung, and what every plan includes.
Video Studio help
A step-by-step walkthrough of the studio, its controls, and what each one changes.
Your first clipcosts nothing
Create a free account, upload a still, and see what 5 seconds of motion looks like before you decide anything else.